Batteries for AMR


New developments in automated meter reading (AMR) technology, such as the advent of automated meter management (AMM) deployments with two-way, realtime communication between utilities and their customers, place an ever increasing demand on the batteries that power the new generation of ‘smart’ meters.

So now, rather than having to supply pulses of just a few milliamps to support radio communication over a few meters, the battery might have to supply a couple of amps to power a GPRS transceiver. It also has to sustain this high level of performance for up to 20 years, even in the most demanding environmental conditions. What then are the key questions a designer must consider to ensure they specify the optimum battery for their particular AMR application?

AMM background

Until quite recently the main role of AMR was seen as providing a communication link that enabled utility meters, such as electricity, water and gas, to be read and billed without the need to gain access to a customer’s premises. These meters are ‘read’ either by a hand-held terminal or by the drive-by method in which the data is received by a mobile receiver in a vehicle. However, a new generation of AMR has now been developed in which the meter is placed in direct, constant two-way communication with the utility. So now rather than using AMR for billing purposes only, the utility can take advantage of AMM as a key tool to help manage its customer relationships by providing complete transparency regarding consumption, tariffs and usage profiles.

1 92 fig 1

One of the key advantages of AMM is of course cost savings. The total annual cost for manual reading of electricity meters in the European Union is estimated at €4.6 billion – or €20 per reading. Italy’s power distribution company ENEL has pioneered the largest and most advanced AMM deployment to date (the Telegestore project completed in 2005), covering some 30 million customers, and estimates that it is now saving €400 million annually.

A further driver for AMM is the increasing liberalisation of energy markets, especially in Europe, combined with tighter legislation for reading of electricity meters that aims to establish clear relationships between electricity consumption and cost. For example, in Denmark, Norway and Sweden the regulator requires hourly meter readings for heavy users of electricity. Sweden will also require monthly readings for all electricity customers from mid-2009.

AMM communicati on architectures

In general there are two competing system architectures for AMM deployments. First, there is the ‘one-stop’ arrangement in which a single meter per customer provides either a low power, short-range radio signal that is read by walk-by/ drive-by or a high power signal carried by the GSM/CPRS mobile communications network. Then there is the ‘twostep’ arrangement in which anything up to 100 meters are networked by hard-wired connection to a data concentrator that provides the two-way communication with the utility control centre, usually by GSM/GPRS. This is the approach adopted by ENEL, with 350,000 concentrators installed in almost every secondary substation.

There are also a number of competing technologies for the communication links between the meters and the utility. Wireless RF and GSM/GPRS that require battery power have been mentioned above. It is also feasible for the AMR signals to be carried by PLC (power line communication) using the power cables themselves for communication, or for a dedicated telephone line to be installed. While neither of these fixed-line methods requires a battery to power the data transmission process, they may use a smaller battery for backup purposes.

Battery selection

Assuming that a wireless AMR system is to be specified, then selection of the correct battery becomes a critical decision. The first question ‘what battery chemistry to use?’ is very simple to answer. The need for high voltage, high capacity, high current pulses and long operating life in demanding operating conditions including extreme temperatures, humidity, dust and UV means that, currently, a primary lithium battery has to be used.

Li-SOCL2 cells

Lithium is the lightest of all metals, and exhibits an exceptional specific capacity (3.86 Ah/gram) and unique electrochemical characteristics. Combining lithium with manganese dioxide (MnO2) powder, polycarbon monofluoride (CFx), or low freezing point liquid cathode materials, such as thionyl chloride (SOCl2) or sulphur dioxide (SO2), results in primary cells with high energy, low weight, reduced self-discharge and the ability to operate under extreme conditions such as temperatures ranging from -40OC to +95OC. Lithium thionyl chloride (Li-SOCl2) cells have a nominal voltage of 3.6 V, while both Li-SO2 and Li-MnO2 cells have a nominal voltage of 3.0 V. Li-CFx cells have a nominal voltage of 2.8 V, and are not normally used in AMR in data transmission applications, but may be used to provide memory backup.

Of the available lithium battery chemistries, Li-SOCl2 cells offer the best choice for AMR applications, since they combine high energy density (typically 1,200 Wh/l), excellent temperature characteristics, low self discharge rates and excellent safety characteristics. The higher cell voltage is a further advantage, since it could enable one battery to be used rather than two lower voltage cells, and ensures compatibility with electronic circuitry which may need a minimum of say, 2.9 V, to function. This is particularly important as the voltage of lithium cells becomes depressed at lower temperatures, so a 3.0 V nominal system could fall below 2.9 V. Li-SOCl2 cells are generally available in a range of sizes from 1/2AA up to DD. Operating temperature A key factor in selecting an AMR battery is the ambient operating temperature, as this can vary considerably according to location. For example, meters in outdoor applications can routinely experience temperatures as low as -30°C, while a meter fixed in a utility room, close to a boiler, can be subjected to constant temperatures of +50°C. While both these temperatures are well within the Li-SOCl2 cell’s normal operating range, prolonged exposure will impinge on its performance and operating life.

Pulse capability

In addition to the nominal battery voltage and cut-off voltage, identifying the pulse profile – cut-off voltage, current, duration and frequency – that has to be sustained is crucial to identifying the right battery for the application. If the AMR utilises low power RF, with a required reading range of just a few metres, then the base current drawn during its ‘sleep’ mode will be just a few microamps, with pulses during data reading and transmission of around 10 mA for a few milliseconds at intervals ranging from hourly to just a few times per day.

The demand on a battery for GSM/GPRS is significantly different, as more power is required to transmit a signal to a receiver several hundred metres distant, or even many thousand of metres if communicating directly with a satellite. In this case the data pulse can range from 500 mA to 2 A for a few milliseconds.

The passivation challenge

Although the Li-SOCl2 cell chemistry offers significant advantages in AMR applications it does have one drawback for high pulse applications, due to a phenomenon known as ‘passivation’. This is because the metallic lithium in the cell rusts, just like iron, when in contact with the liquid thionyl chloride, producing a thin ‘passivation’ layer that protects the lithium from further reaction. Under normal conditions, this layer does not degrade cell performance. Indeed, it performs a very useful function as it prevents major loss of the cell capacity, known as ‘self-discharge’, while the cell is in storage, resulting in a long shelf-life. It also contributes to a long service life.

However, should the passivation layer grow too thick, then the cell’s discharge performance may be affected. The growth of the layer is influenced greatly by storage conditions, and long periods of inactivity at elevated operating temperatures will cause it to grow in thickness. A passivated cell may exhibit voltage delay, which is the time lag that occurs between the application of a load on the cell and the voltage response.

As the passivation layer thickens, the voltage delay becomes more severe, and this could affect the operation of the AMR system. An important element of the expertise of the battery manufacturer, and a strong differentiation factor regarding the in-service performance of the product, lies in the correct selection of the special additives that may be added to the cell’s liquid electrolyte to regulate passivation. These additives result in the formation of a useful protection layer during storage, which peels off swiftly when current pulses are applied.

There are also two other possible solutions to passivation. One is to work with the designer of the AMR system to implement special protocols that reduce the periods of cell inactivity by calling for more regular high current pulses, hence reducing the growth of the passivation layer. The second approach is to fit a capacitor that is charged by the battery in order to power the high current pulses. It is recharged by the battery, in advance of the next pulse, and therefore eliminates passivation effects.

Life time modelling

As can be seen making the right choice of battery calls for the AMR designer to work in partnership with the battery supplier to ensure that all the key factors – base current, pulse currents, cut-off voltage, temperature and environmental conditions – are considered. This is especially important as most AMR manufacturers require a battery service life of at least five years, while many call for over 10 years and some now look for a battery that will last as long as their meter – 15 to 20 years.

Field and laboratory data collected over 30 years by Saft have been developed into a life time model that enables the expected life of primary cells in this demanding application to be predicted accurately by evaluating the specific utilisation profile. Calculated results are then combined with results from long-term bench tests to produce the most accurate life prediction possible.